Linear infrastructure drives biotic homogenization among bird species of a tropical dry forest

Linear infrastructures (LIs) such as roads, railroads, and powerlines are expanding rapidly around the globe. While most future developments are projected to take place in tropical regions, available information on impacts of LIs is biased towards single species studies of solely road impacts in temperate regions. Therefore, we investigated impacts of three types of LIs (road, railroad, and powerline) on the bird community of a tropical dry forest. Point-count surveys to record avian richness and abundance were conducted at 80 plots that were spatially stratified to include sites proximate to all possible LI combinations. Five measures of vegetation structure were collected at each plot as well. We then assessed the relationship between the bird community (i.e., richness, abundance, composition) and distance to each LI type while accounting for variation in vegetation structure. Species richness and abundance both declined significantly (25% and 20%, respectively) from edge habitat next to railroad to interior forest plots, while community composition was significantly altered by the distance to all three LIs. Road and railroad (both forms of dynamic infrastructure with moving vehicles) had similar effects on the bird community that contrasted with those of powerline (a type of static infrastructure). The resulting ordination reveled that Sri Lankan endemics are significantly disfavored by LI proximity, while species that now have naturalized populations across the globe are most often found proximal to LI. Our results emphasize that LI drive biotic homogenization by favoring these increasingly widespread species at the expense of unique elements of the biota.

Methods
Study Site

The study was conducted in the Polonnaruwa District, North Central Province, Sri Lanka. The landscape consists of two protected areas: Thalapathkanda Forest Reserve and Minneriya National Park. The region has all three LI types (road, railroad, powerline) in close proximity and is scheduled for infrastructure expansion.

Bird Survey Method

Point-count surveys conducted at 80 spatially stratified points.

Distance Categories: Plots were categorized based on their proximity to LI.

Survey Duration: Each survey lasted 20 minutes and was conducted between 0530 - 0800 h.

Survey Seasons:

Dry season: July - August 2019

Wet season: December 2019 - January 2020

Dry season: July - August 2020

Vegetation Structure Measurements

Canopy cover: Percent canopy cover measured using ImageJ software.

Shrub cover: Percentage of understory vegetation cover measured with a white background board.

Herb cover: Percentage of herbaceous cover measured via quadrats.

Diameter at Breast Height (DBH): Count of trees with DBH >10cm.

Plant species richness: Number of tree species within each plot.

Data Analysis

Impact on Species Richness & Abundance: General linear models with AICc selection.

Community Composition: Distance-based redundancy analysis (db-RDA).

Software Used: R ver 3.6.0, JMP Pro 15.